OpenAI and Anthropic Call for UK Legislative AI Regulation; Google and Meta Refuse to Endorse
nashnova research
OpenAI and Anthropic told a UK parliamentary committee they support legally binding rules for frontier AI; Google DeepMind and Meta refused to endorse the idea — splitting the industry's biggest players into two open camps on whether governments should regulate them.
Who backed legislation, and who dodged?
Anthropic's Europe lead Liam Booth-Smith was explicit: "Voluntary commitments are not enough — companies need binding rules."
OpenAI VP Tom Duff Gordon was more measured but pointed the same way — he backed a statutory framework and international conventions, suggesting mandatory incident reporting as a starting point.
This means → both companies actively asked to be regulated, closing off the voluntary-compliance escape route.
Why did Google and Meta stay silent?
Google DeepMind's Owen Larter, asked directly whether he supported a statutory framework, replied: "We will focus on standards" — sidestepping the legislation question.
Meta's Rob Sherman pointed to existing UK regulators, saying "we are not writing on a blank page" — implying current rules are sufficient.
In plain terms = neither company said "no," but neither said "yes." In a parliamentary record, that evasion is itself a position.
Why did Parliament want this split on the record?
Committee chair Liam Byrne said the goal was to "expose the huge divergences among some of the richest companies on earth" and prove that "voluntary self-regulation cannot be relied upon."
The hearing was cross-party; alongside the four companies' executives, it called four former-employee whistleblowers and senior civil servants.
This reflects a deliberate strategy: make the split public first, then use it as political ammunition to push legislation.
Why did MI5's warning land on the same day?
On the same day, MI5 Director General Ken McCallum warned in his annual threat assessment that AI misuse in biological weapons is his top concern.
His words: the possibility that AI could reshape the bio-weapons risk landscape is "the misuse scenario to which I attach the greatest weight" — even though it has not yet materialised.
This means → a national-security warning arriving alongside the parliamentary hearing created a twin narrative — industry split + security alarm — that sharpens the case for urgent legislation.
What to watch next?
OpenAI and Anthropic's UK stance tracks their partial US positioning, but this is the first time the split with Google and Meta has been entered into a parliamentary record.
Put simply = the regulatory-path dispute has moved from an internal industry debate to an open legislative contest.
Whether the UK moves first on binding AI legislation is the key test of whether this split translates into actual policy change.
市场有风险，内容仅供研究参考，不构成投资建议。